Bệnh học đốt sống ppsx

Percutaneous Treatment of Vertebral Body Pathology Jeffrey M. Spivak, MD, and Michael G. Johnson, MD Abstract Minimally invasive methods of treating the symptoms of spinal compression fractures have attracted the attention and interest of both surgeons and patients over the past two de- cades. Vertebroplasty, the percutaneous augmentation of vertebral body volume and strength by injection of a hardening material in a liquid state, was originally used to manage painful neoplastic vertebral body lesions and pathologic fractures. Surgical acrylic cement was injected during open surgery to strengthen vertebral bodies with angiomas. 1 An analgesic effect was noted, and indications for the techniquein neoplasticdisease ex- panded to include both primary and secondary pathology. Although most metastatic lesions of the spine are effectively managed with radiation therapy, pain relief is not universally successful for those with marked structural compromise or collapse. Vertebroplasty provides both analgesia and vertebral structural stabilization in a single procedure. Kyphoplasty notonly can facilitate increased stability, but it also may allow some correction in cases of acute and subacute fracture collapse and kyphotic deformity. Apercutane- ous inflatable balloon device (bone tamp), approved by the US Food and Drug Administration (FDA) since 1998, is used to restore vertebral body height. The created void is then filled with a low-pressure, high-viscosity injection of a hardening support material. Polymethylmethacrylate (PMMA) is currently used in both procedures. Both vertebroplasty and kyphoplasty can be used to treat painful acute, subacute, and chronic osteoporotic compression fractures caused by decreased bone mineral density and changes in bone ar chitecture that accompany senile osteoporosis, espe- cially in postmenopausal older women. 2-4 It is estimated that age- related osteoporotic compression fractures occur in more than 500,000 patients per year in the United States. 5 Most of these fractures are successfully managed with analgesic medications and orthoses. Medical evaluation and pharmacologic management of a patient’s underlying osteoporosis is imperative to minimize the risk of additional vertebral fracture. Physical exercises, preferably movement exercises and extension protocols, are instituted once the initial pain has resolved. Only rarely do these fractures involve the posterior vertebral cortex and result in spinal canal compromise and neurologic deficit. Dr. Spivak is Director, The Hospital for Joint Dis- eases Spine Center, New York, NY. Dr. Johnson is Assistant Professor, Orthopaedics and Neuro- surgery, University of Manitoba, Manitoba, Win- nipeg, Canada. None of the following authors or the departments with which they are affiliated has received anything of value from or owns stock in a commercial com- pany or institution related directly or indirectly to the subject of this article: Dr. Spivak and Dr. Johnson. Reprint requests: Dr. Spivak, The Hospital for Joint Diseases Spine Center, 301 East 17th Street, New York, NY 10003. Copyright 2005 by the American Academy of Orthopaedic Surgeons. Percutaneous vertebral body injection procedures currently are used to stabilize and reinforce weakened or fractured bone resulting from metastatic disease and severe osteoporosis. Both vertebroplasty and kyphoplasty can reinforce the structure of a vertebral body and provide pain relief, but the procedures have technical differences. Kyphoplasty improves vertebral height to varying degrees in nearly three quarters of patients. Kyphosis is improved more effectively when the procedure is performed within 3 months from the onset of fracture pain. To date, it is unknown whether vertebroplasty with preprocedure postural reduction can provide similar improvement of deformity. Complications are relatively infrequent with both vertebroplasty and kyphoplasty. Cement leakage from the vertebral body is more likely with vertebroplasty than with kyphoplasty. Leakage is more common in the treatment of pathologic fractures resulting from metastatic disease. Clinical complications caused by cement leakage and neural compression are infrequent. Specific indications for these injection procedures need to be more clearly refined. Long-term outcomes, including the fate of the injected material and the effect on adjacent vertebrae, have yet to be determined. J Am Acad Orthop Surg 2005;13:6-17 Perspectives on Modern Orthopaedics 6 Journal of the American Academy of Orthopaedic Surgeons Although many physicians think that osteoporotic compression fractures routinely heal uneventfully without significant long-term se- quelae, this may not be accurate. Both thoracic and lumbar compression fractures have been associated with decreased lung capacity. 6 The pres- ence of five or more thoracic compression fractures has been associated with a higher risk of death from pulmonary complications in women aged 65 years and older. 7 Residual kyphotic deformity may r esult in loss of sagittal balance with significant limitation of functional abilities, pain, and a generalized poor overall health perception. 8 In other cases, a pro- longed period of incapacitating back pain affects the patient’s quality of life and increases the likelihood of bed rest–related morbidity, including pressure sores, pneumonia, and thromboembolic disease. Indications Both vertebroplasty and kyphoplasty may be used to augment painful weakened or fractured vertebrae in a variety of clinical settings (Table 1). Controversy exists as to the specific indications for these procedur es. Pro- ponents of their aggressive early use are generally concerned with the potential disease-related morbidity, including progressive kyphosis and persistent pain and disability. Advocates of more limited use ar e concerned with the incidence of procedure-related morbidity, including immediate complications, the long-term effects of the cement used, and potential negative biomechanical effects on adjacent vertebrae. Stabilization of a painful metastatic vertebral lesion is the least controversial indication for percutaneous vertebral augmentation. Interruption of the posterior vertebral cortex by tumor or fracture is a relative contraindication because risk of extravasation into the spinal canal is increased. Sig- nificant spinal canal compromise by tumor or fracture fragments is an absolute contraindication, as is complete vertebral body collapse. Currently, both vertebr oplasty and kyphoplasty are most commonly used to stabilize acute and subacute osteoporotic vertebral compression fractures with collapse and/or kyphosis (Fig. 1). With vertebroplasty, some positional reduction of the kyphotic deformity may be achieved. Although the low-viscosity cement injection will not further correct any kyphosis, generally it will limit progression. Kyphoplasty has the add- ed theoretical benefit of being able to correct kyphosis because insufflation of the bone tamp can elevate the de- pressed end plates. Therefore, kyphoplasty may be preferable when correction of kyphosis is a primary indication. The medical condition of the patient must be taken into consideration when selecting one of these invasive procedures. Kyphoplasty typically is performed under general anesthesia, whereas vertebroplasty routinely is done using only intravenous sedation and local anesthetic. Therefore, vertebroplasty may be preferred in patients with multiple significant medical comorbidities who are being treated for painful vertebral lesions with minimal or no deformity. Patient Assessment Clinical assessment of the patient with a painful vertebral lesion or fracture includes a complete history and physical examination along with appropriate radiologic studies. Important in- formation includes the nature of any trauma associated with the onset of pain, known occurrence of a primary or metastatic lesion, previous osteoporotic fractures, and an assessment of risk factors for osteoporosis if that diagnosis has not yet been made. Physical examination should include measurement of height compared with historic data from the patient, visualization of thoracic kyphosis or loss of lumbar lordosis, palpation of tenderness along spinous processes Table 1 Indications and Contraindications for Percutaneous Vertebral Augmentation by Vertebroplasty or Kyphoplasty Indications Contraindications Absolute Absolute Painful metastatic lesion with intact posterior vertebral cortex Fractures or neoplasms with spinal canal compromise Chronic (>3 months) compression fracture with nonunion Vertebra plana (complete vertebral body collapse) Progressive kyphosis to ≥20° in a subacute (<3 months) compression fracture Relative Relative Subacute (<3 months) compression fracture with persistent pain and dysfunction despite adequate nonsurgical management Fractures and metastatic vertebrae with posterior vertebral cortex involvement and compromise Acute compression fracture with ≥20° kyphosis or ≥40% collapse Less than one-third vertebral body height remaining Acute compression fracture with one or more prior compression fractures and kyphotic deformity Jeffrey M. Spivak, MD, and Michael G. Johnson, MD Vol 13, No 1, January/February 2005 7 or widening between adjacent pr ocess- es, and a thorough neurologic evaluation. Plain radiographs should include standing anteroposterior (AP) and lateral views centered at the level of the fractured vertebra, as well as a standing lateral 36-in radiograph of the entire spine to assess overall re- gional and segmental sagittal align- ment. Flexion and extension lateral radiographs may be helpful in assessing the degree of fracture mobility and healing (Fig. 2). If tolerated by the patient, a hyperextension lateral radiograph over a bolster should be considered before performing an injection procedure, to assess the potential for postural r eduction of vertebral collapse and kyphosis. Magnetic resonance imaging (MRI) is able to visualize both vertebral and soft-tissue pathology. In patients with osteopor otic fractures, diffuse or focal signal changes of decreased intensity on T1-weighted images and increased signal intensity on T2-weighted images may be in- dicative of an active fracture not yet fully united. MRI is particularly helpful with multiple compression deformities when trying to determine which are still active and therefore more likely to be painful. The extent of marrow involvement in patients with metastatic disease can be as- sessed, with special attention paid to the nonfractured vertebrae. MRI also will demonstrate the soft-tissue mass associated with a metastatic pathologic fracture and can be used to assess the integrity of the posterior vertebral cortex. The integrity of the posterior wall may not be as well visualized as with computed tomogra- phy (CT) scans; therefore, any ques- tionable involvement or compromise of the posterior vertebral cortex seen on MRI should prompt further evaluation with a CT scan. CT provides excellent detail of the bone involvement, and it is the best imaging procedure for assessing the extent of vertebral body and posterior element fracture. The combina- tion of CT scan and plain radiographs is helpful to fully classify the fracture type. It is important to distinguish between a compression fracture, with collapse of the anterior vertebral cortex (with or without upper or lower end plate involvement), and a burst fracture, in which the posterior wall of the vertebra is fractured, as well. 9 Generally, burst fractures demonstrate more uniform vertebral collapse, whereas compression fractures show more anterior wedging or kyphotic collapse. Technetium Tc 99m–labeled bone scanning can be useful in assessing for increased metabolic activity de- noting increased bone turnover, as seen in infiltrative lesions and incom- pletely healed fractures. Bone scans are particularly useful in assessing Figure 1 A, Lateral radiograph in a 72-year-old man made 2 days after a minor fall demonstrating a compression fracture of L1. B, Lateral radiograph taken 3 weeks later demonstrating progression of anterior vertebral collapse at a time when the patient was having worsening pain. The patient was treated with kyphoplasty. Postoperative lateral (C) and anteroposterior (D) radiographs showing restoration of the collapse and minor disk space extravasation of cement. The patient had immediate full resolution of pain, which persisted at 6 months postoperatively. Percutaneous Treatment of Vertebral Body Pathology 8 Journal of the American Academy of Orthopaedic Surgeons fracture activity in patients with multiple compression deformities. In- creased vertebral activity on bone scan has been shown to be highly pre- dictive of a positive clinical response to vertebroplasty. 10 Technique Vertebroplasty and kyphoplasty are both done with the patient positioned prone on a radiolucent table. This al- lows for the use of biplanar fluoros- copy. The use of a Jackson radiolucent table and dual image intensifiers (which provide simultaneous posteroanterior [PA] and lateral imaging) is optimal. However, a single image intensifier can be used, rotating between PA and lateral images as needed. For vertebroplasty, local anesthetic is combined with intravenous sedation; for kyphoplasty, a general anesthetic is more commonly used. Two techniques, the transpedicular and the extrapedicular, can be used to place the injection trocar or working cannula. 11 The transpedicular approach is more commonly used in the lumbar spine (Fig. 3). The tip of the injection trocar is started at the upper/outer corner of the visualized pedicle outline as seen on the PA radiograph (left pedicle, 10 o’clock position; right pedicle, 2 o’clock position). Proper overall sagittal trajectory and starting point within the superior-inferior pedicle border are confirmed on the lateral radiograph before trocar advancement. Trocar advancement is visualized fluoroscop- Figure 2 A 77-year-old man presented with a 4-month history of unremitting midback pain that limited ambulation and functional ability and required narcotic use. A, Standing lateral radiograph demonstrating an anterior compression deformity of T9. Note the diffuse spondy- losis and multilevel anterior bridging osteophytes. B, Supine extension lateral radiograph demonstrating an intervertebral defect (arrow) just below the superior end plate. C, T2-weighted sagittal MRI scan demonstrating the fluid-filled intervertebral cleft characteristic of a fracture nonunion. Anteroposterior (D) and lateral (E) intraoperative fluoroscopic views with bone tamps inflated. F, Postoperative lateral radiograph after stabilization using percutaneous kyphoplasty. Note the small amount of clinically asymptomatic anterior extravasation. The patient experienced complete pain relief and remained asymptomatic 1 year postoperatively. Jeffrey M. Spivak, MD, and Michael G. Johnson, MD Vol 13, No 1, January/February 2005 9 ically. When the trocar tip is midway along the length of the pedicle on the lateral view, it should be central in the pedicle outline on the PA view. When it is through the pedicle and at the posterior vertebral cortical mar gin on the lateral view, the trocar should be just within the medial border of the pedicle outline on the PA view. For vertebroplasty, the trocar is advanced until the tip is at the junction of the anterior and middle thirds of the vertebral body. For kyphoplasty, the positioning trocar is exchanged for a working cannula over a guidewire. The cannula is positioned near the posterior vertebral body margin while the working instruments (ie, drill, bone tamp, cement inserters) are advanced anteriorly until they are ap- proximately 3 mm from the anterior vertebral body border. Convergence of the tip toward midline should be followed on the PA view. The extrapedicular approach is commonly used in the thoracic spine (Fig. 4). The starting point is more lateral than the pedicle, with the trajectory more medially directed, which enables a more medial trocar tip placement within the vertebral body than a transpedicular route through the sagittally oriented thoracic pedi- cles would allow. The trocar tip i s first inserted lateral to the pedicle, either through the thoracic transverse process or along the transverse process/ rib junction. Under biplanar fluoroscopic visualization, the needle or trocar is advanced along the medial border of the rib until the lateral border of the pedicle is reached on the PA view and the posterior vertebral body margin is reached on the lateral image. For kyphoplasty, the trocar Figure 3 Trajectory of the transpedicular approach. A, Posteroanterior view. B, Lateral view. C, Axial representation. Symbols along the trajectory indicate the position of the trocar tip at various depths of insertion. The tip should not pass medial to the medial border of the pedicle on the posteroanterior view until it is anterior to the posterior margin of the vertebral body on the lateral view. = insertion point, ◊ = point at pedicle/vertebral body junction, = midvertebral body. Figure 4 Trajectory of the extrapedicular approach for needle or trocar placement. A, Posteroanterior view. B, Lateral view. C, Axial representation. Symbols along the trajectory indicate the position of the trocar tip at various depths of insertion. The tip should not pass medial to the lateral border of the pedicle on the posteroanterior view before the posterior aspect of the vertebral body is seen on the lateral view. = insertion point, ◊ = point at pedicle/vertebral body junction, = midvertebral body. Percutaneous Treatment of Vertebral Body Pathology 10 Journal of the American Academy of Orthopaedic Surgeons is advanced only minimally beyond this position until the tip is anterior to the posterior aspect of the vertebral body on the PAview and just medial to the lateral border of the pedicle on the lateral view. The trocar is then exchanged for the working cannula over a guidewire, with further advancement of the working instruments beyond the cannula toward the anterior vertebral body. For vertebroplasty, the injection trocar is advanced further anteromedially toward the central aspect of the vertebral body. An 11-gauge injection trocar is used for vertebr oplasty. Regardless of the trocar placement approach, intra- vertebral venography is performed before injecting the PMMAto exclude the placement of the needle tip within a major vascular channel. Avenous blush within the vertebral body indicates proper needle placement because the injected dye must remain and pass through the vertebral marrow before exiting through the venous sinusoids. Lack of a venous blush indicates rapid release of the dye directly out of the vertebra on injection, so the trocar tip should be re- positioned. Once proper trocar position is confirmed, the cement is mixed and injected into the vertebral body in a low viscosity state under contin- uous fluoroscopic visualization. Injec- tion is continued until hemivertebral or holovertebral filling is achieved and no more cement can be delivered into the body. It is stopped immediately if any extravasation is noted into the surrounding veins, the spinal canal posterior to the vertebral body, or the disk space. Clinical improvement has not been shown to correlate with the amount of cement injected. After completing the injection, the trocar is removed, and hemostasis is obtained by pressure. The contralateral hemi- vertebra is treated in the same man- ner when a bilateral injection is to be used. Kyphoplasty also can be performed percutaneously through either a transpedicular or lateral extrapedicular approach. Bilateral 3.5 mm– diameter working cannulas inserted over the 11-gauge positioning trocars are used to create the working channels within the vertebra using a hand drill. A hollow trocar often is used in place of one of the drills to remove a bone biopsy core of tissue. The created channels allow passage of the de- flated bone tamps, which are inflated with a pressurized radio-opaque liquid after being properly positioned. In theory, inflation of the balloons increases the vertical height of the vertebral body, thereby reducing kyphosis. As the bone tamp is inflated, the surrounding cancellous fracture fragments are compacted along the margin of a central void that is created. In one clinical series, the average amount of balloon inflation volume was 2.6 mL (range, 0.5 to 5.0 mL), and the average balloon inflation pressure was 130 psi (range, 70 to 250 psi). 12 The balloons are removed, and the defects are simultaneously filled with PMMA inserted manually through the trocars in a relatively high viscosity state using a low injection pressure. By using low injection pressure and a higher viscosity cement while injecting into a bony void surround- ed by compacted cancellous bone, this technique theoretically reduces the risk of intravascular injection as well as leakage into either the surrounding tissues or the spinal canal. An in vivo study of contrast injection leakage before and after balloon inflation and fracture reduction sup- ports this theory. The authors noted significantly less vascular and trans- cortical extravasation of contrast after use of the inflatable bone tamps (P = 0.001 for each). 13 Postoperative Management Patients are mobilized as soon as the sedation or general anesthetic has worn off. No period of bed rest is needed. Many patients have restrict- ed activity for weeks or longer because of their painful fractures and may be relatively debilitated both before and immediately after the injection procedure. Progressive return to full activities is recommended, without specific r estrictions. Physical therapy for gait training, aerobic condi- tioning, and back-specific stretching and strengthening may benefit many of these patients. Most patients do not require assistive devices for ambulation, but such devices may be useful for patients who have been nonam- bulatory because of their fractures. Postoperative bracing is not used routinely because the fractures are ren- dered mechanically stable by the cement injection. Results The outcome of vertebral injection procedures is highly dependent on the clinical indication for the procedure and, to a lesser extent, the technique used. Most published series of vertebroplasty have reported on its use to treat patients with neoplastic lesions and/or osteopor otic compression fractures. 6,11,14-17 More recent kyphoplasty series have reported only on patients with osteoporotic compression fractures. 18,19 Patients with painful metastatic vertebral lesions have had good clinical results with vertebroplasty. 16,17,20 In one series of 37 patients with vertebral metastases, 73% had marked clinical improvement after percutaneous vertebroplasty. 17 Improvement was defined as a decrease in pain resulting in the dose of analgesia being reduced by 50% or as a change from a narcotic to a nonnarcotic medica- tion. The decrease in pain after vertebroplasty was maintained at follow- up, with 76% of patients reporting persistent pain relief at 6 months and 65% at 1 year. Another series reported excellent pain relief in 11 of 18 metastatic lesions. 16 The pain relief experienced by patients undergoing vertebroplasty is Jeffrey M. Spivak, MD, and Michael G. Johnson, MD Vol 13, No 1, January/February 2005 11 similar to that with radiation therapy for painful metastatic lesions of bone. 21,22 However, vertebroplasty is associated with a more rapid improvement in pain, with 80% of patients noting impr ovement within the first 24 hours and being able to stand the day after the pr ocedure. 23 Pain relief normally is seen 2 to 10 days after nonfractionated radiation therapy and 1 to 2 weeks after conventional radiation therapy. 21,22 Vertebroplasty is not an alternative to radiation therapy; rather, it can be used as a com- plementary adjunct procedure. Ce- ment injection provides a mechanical pain treatment by stabilizing the spinal segment; radiation therapy provides a biologic treatment modality. The mechanism by which vertebroplasty with injection of PMMAre- lieves pain is not clear. Neoplasms are not directly innervated; rather, the pain is thought to be caused by fracture or impending fracture stressing of the remaining bone. Nerve endings in the r emaining normal bone may be stimulated by the mass effect of the neoplastic tissue. Paraneoplastic production of humoral mediators also may play a role in pain production. The analgesic effect of vertebral injection is thought to be the result of im- mobilization of the bone and structural support and possibly of the destruction of the terminal nerve endings by the cytotoxic or thermal effects of PMMA. The relationship between amount of vertebral fill fr om percutaneous vertebroplasty using PMMAand pain reduction has been examined in a series of 40 cases (37 patients) of metastatic disease and myeloma using postinjection CT scans. 15 Lesion filling was >75% in only 5 of the 40 cases, 50% to 74% in 14, 25% to 49% in 13, and <25% in 8. No statistically significant relationship was found between the cause of the osteolytic lesion, the percentage of vertebral body filling by PMMA, and pain reduction. Several clinical series have reported results of vertebroplasty for painful osteoporotic vertebral compr ession fractures, with the percentage of good and excellent r esults ranging fr om 75% to 100%. 11,14,16,19,20,24-26 In general, however, these series are retr ospective, with limited duration of follow-up. Jensen et al 11 reported on 29 patients who underwent percutaneous vertebr oplasty to manage osteoporotic vertebral body compression fractures. Two patients had complete resolution of back pain immediately after vertebroplasty. Twenty-six patients (90%) described pain relief and improved mobility within 24 hours of treatment. All patients who had required parenteral narcotics tolerated reduction to oral medications. Three patients had no significant pain relief and were continued on pretreatment medical reg- imens. 11 Another prospective series of 16 patients undergoing vertebroplasty for painful osteoporotic compression fractures showed a significant (P < 0.005) decrease in pain by day 3 postprocedure, which was maintained at 6 months. 14 Asingle series of long-term (mean, 48 months) results after vertebroplasty for osteoporotic fractures showed maintenance of pain relief at long-term follow-up similar to that achieved 1 month after the procedure. 27 A preliminary report of a multi- center study of kyphoplasty for osteoporotic compression fractures in 340 patients (603 fracture levels) found a 90% initial symptomatic and functional improvement rate and significant (P < 0.05) restoration of the anterior, middle, and posterior vertebral body height when performed within 3 months of the onset of the fracture. 18 In a prospective study of 70 consec- utive kyphoplasty levels in 30 patients, 12 the indication for kyphoplasty was a painful osteoporotic compression fracture (mean duration of symptoms, 5.9 months). Clinical assessment, using preoperative and latest postoperative data from the Medical Out- comes Study 36-Item Short Form, showed significant (P < 0.01) improvement in scores for bodily pain, physical function, vitality, mental health, and social functioning at a mean follow-up of 6.7 months. Only general health and role emotional scores did not show significant improvement. Radiographic evaluation before and after kyphoplasty indicated varying amounts of height restoration in 70% of the levels injected and no height restoration in 30%. Overall height loss by the fracture averaged 8.7 mm (range, 2 to 17 mm). Overall height loss restoration among all injected levels was 2.9 mm (an average of 35%). However, in the levels showing any height restoration (70% of total levels), the average height restoration was 4.1 mm (47% of lost height). Ce- ment leakage was seen at six levels (8.6%): twice into the disk space, three times into the soft tissues, and once into the epidural space. None of these incidents was clinically significant. 12 Complications Reported complication rates for vertebroplasty have been very low in nu- merous series. 1,15,17,25,28,29 In decreas- ing incidence, r eported complications include cement leakage, which can range from asymptomatic to causing neurologic compromise; cerebrospi- nal fluid leak; cement embolization causing pulmonary embolism; and infection (Fig. 5). PMMA leakage was documented in 29 of 40 cases by CT scan evaluation after percutaneous vertebroplasty for metastatic lesions and myeloma. 15 Fifteen of the leaks were into the spinal canal, 8 into the neural foramina, 8 into the adjacent disk space, 21 into the paravertebral tissue, and 2 into the lumbar venous plexus. Ce- ment leakage into the epidural space was strongly associated with posterior vertebral cortical destruction by fracture and tumor. Leakage of PMMA into the neural foramina was associated with posterior cortical destruction in six of the eight cases. Intradis- kal PMMA leakage was associated Percutaneous Treatment of Vertebral Body Pathology 12 Journal of the American Academy of Orthopaedic Surgeons with cortical fracture or osteolysis of the vertebral end plates in all cases. None of the documented leaks com- promised axial pain relief. One leak into the psoas muscle produced a transient femoral neur opathy that r esolved in 3 days. One of the foraminal leaks produced nerve root compression requiring decompressive surgery 1 month after the procedure. In another clinical series of 53 injected levels in 35 patients, the overall complication rate was 6% per level treated, with two cases of cement leakage requiring delayed surgical inter- vention. 29 Three other asymptomatic cement leakages were noted, and one additional patient had a suspected ce- rebrospinal fluid leak requiring 1 week of hospitalization and hydra- tion. Proper needle placement, PMMA opacification, and real-time fluoroscopic injection visualization help minimize these technical complications. Technique-related vertebroplasty complications can be minimized by strict adherence to specific patient in- clusion and exclusion criteria. Weill et al 17 requir ed preservation of at least one third of the vertebral body height before performing vertebroplasty. Al- though these authors did not consid- er posterior vertebral wall destruction to be an absolute contraindication (even though there is a documented increased risk of extrusion of the cement or soft tissue into the spinal canal), others would disagree. The risk of cord compr ession in the cervical and thoracic levels is relatively high, and the decision to perform vertebroplasty should be made only after careful consideration of the potential for alternative treatments. In their series of 37 patients with spinal metastases, 40% of whom had varying degr ees of posterior vertebral body wall destruction, Weill et al 17 reported no cases of significant cord compression. There was only one case of postprocedure radicu- lopathy caused by nerve r oot compression that required surgery. There was a 38% incidence of cement leakage toward the disk, epidural fat, perivertebral soft tissue, epidural veins, and perivertebral veins. The leaks were symptomatic in five cases. Two leaks into the vena cava wer e reported; however, no clinically significant complications resulted. 17 Complications reported with kyphoplasty include cement extravasation as well as rib fractures, which in one study occurred in 2 of 3 0 patients secondary to patient positioning. 12 One case of pulmonary edema and myocardial infarction was attributed to intraoperative fluid overload. Oth- er reported complications include transient fever, hypoxia, and postoperative epidural hematoma (associated with a heparin bolus 8 hours postoperatively). Garfin et al 18 reported two cases of neurologic injury secondary to problems with needle insertion and positioning of the cement filling tube with epidural cement extravasation. Intraoperative balloon rupture occurred 14 times, chiefly at the end of inflation. All broken bone tamps were easily removed, and three cases required reinsertion of a new balloon to complete the inflation. No direct patient complications resulted from balloon failure. Postinjection pulmonary embolism is rare; it may be related to low viscosity cement on injection or to leakage of the methacrylate monomer. 29-31 Acute hypotension and incr eased Pco 2 Figure 5 Images after vertebroplasty done for a pathologic metastatic vertebral fracture of L1 in a 73-year-old woman. Lateral (A) and anteroposterior (B) radiographs demonstrating cement extravasation into the disk spaces laterally and along the needle tract. C, Axial CT image demonstrating cement leakage into the left pedicle tract (straight arrow) and the right epidural space (curved arrow). The patient became paraplegic within 2 weeks of the vertebroplasty because of the cement extravasation and tumor growth and required emergent posterior decompression and stabilization. Jeffrey M. Spivak, MD, and Michael G. Johnson, MD Vol 13, No 1, January/February 2005 13 measurements because of fat and marrow element embolism have been reported after cement vertebr oplasty injection into intact sheep vertebrae in vivo. 32 No specific direct, major neu- rotoxic or thermal-related complications of PMMA have been reported. Only one case of infection (in an immunocompromised patient) has been reported in all published clinical series to date. In an attempt to decrease the potential for infection, some surgeons add tobramycin to the PMMA when treating immunocompromised patients. 11 An important theoretical concern after the procedure is the possibility of increased force on the vertebrae adjacent to a cement-injected level, with an increase in adjacent segment fracture rate (Fig. 6). Recent series have examined the incidence of adjacent and remote vertebral fractures after percutaneous cement injection for osteoporotic compression fracture. 27,33 In a series of 25 patients with a mean follow-up of 48 months after vertebroplasty, 13 patients (52%) developed at least one new fracture during the follow-up period, with a total of 34 new fractures. 27 The authors reported a small but significantly increased risk of vertebral fracture in vertebrae adjacent to the cement injection with an odds ratio (OR) of 2.27, compared with an OR of 1.44 for vertebrae adjacent to a fractured, nonce- mented vertebra. Another retrospective review reported that 22 of 177 patients (12.4%) developed a total of 36 new fractures after vertebroplasty. 34 Twenty-four of the 36 new fractures (67%) occurred in adjacent vertebral levels. Another finding of concern was that 24 of the 36 new fractures (67%) occurred within 30 days of the vertebroplasty. 34 A recent retrospective review of 115 patients (225 treated levels) treated by kyphoplasty (average follow-up, 10 months) reported a new fracture incidence of 19% (22/1 15 patients). 33 Sev- enteen of the 22 patients with new fractures had adjacent segment fractures. Fifteen of the 22 r efractur e patients had an initial diagnosis of secondary osteoporosis caused by chronic cortico- steroid use, making this a patient population at high risk of refracture. Refracture was seen in 7 o f 81 primary osteoporosis patients, for an incidence of 8.6% (yearly incidence, 10.3%). 33 These reported refracture rates may not differ significantly from the natural history of the disease for osteoporotic compression fractures. In a series of 2,725 postmenopausal women (mean age, 74 years) followed for 3 years, the incidence of new vertebral fractures in the first year was 6.6%. 35 Of the 381 participants who developed a fracture (a group similar to the large series), the incidence of a new vertebral fracture within the next year increased to 19.2%. Although overall r eported complication rates are low in published series for both vertebroplasty and kyphoplasty, with the increasingly widespread use of these procedures, the true complication rates may be significantly higher. Larger reported Figure 6 An 83-year-old man underwent percutaneous vertebroplasty for a painful L3 fracture. He had initial relief of pain, followed by new acute back pain caused by an L2 fracture with kyphotic collapse. A, Lateral radiograph 6 months after percutaneous vertebroplasty showing the newly fractured L2 level with previous vertebroplasty of L3. The small amount of cement extravasation anterior to L3 and into the L2-3 disk space was of no clinical sig- nificance. B, T2-weighted sagittal MRI scan. Because of progressive pain and kyphotic collapse, the patient underwent percutaneous kyphoplasty with excellent restoration of vertebral height, corr ection of kyphosis, and relief of back pain. C, Intraoperative lateral radiograph demonstrating inflated bone tamps. D, Lateral radiograph with cement injected. The patient had marked and sustained pain improvement over 1 year postoperatively without refracture. (Courtesy of Frank Schwab, MD, New York, NY.) Percutaneous Treatment of Vertebral Body Pathology 14 Journal of the American Academy of Orthopaedic Surgeons series are needed to determine the true safety of these procedures. Experimental Data Several studies have explor ed the biomechanical effects of cement injection into intact and fractured vertebrae. In one study of bipedicular injections into compression-fractured lumbar vertebrae, three different commer- cially available bone cements restored vertebral body stiffness to greater than prefracture values. 36 However, only Simplex P (Howmedica, Rutherford, NJ) and Osteobond (Zimmer, Warsaw, IN) restored vertebral stiffness to initial values. Cranioplastic (Johnson & Johnson, Raynham, MA) did not. In another study of bipedicular injections, α-BSM (bone substitute material; ETEX, Cambridge, MA), a bio- degradable calcium phosphate bone cement, compared favorably with PMMA in strengthening intact and fractured osteoporotic cadaveric vertebral bodies. 37 With either calcium phosphate bone substitute or PMMA augmentation, fracture strength was significantly (P < 0.05) stronger than in the intact control group. Vertebral stiffness after both calcium phosphate bone cement and PMMA augmentation also was significantly (P < 0.05) higher than in the intact control group. Similar studies have reported the efficacy of other bioactive cements. 38,39 Tohmeh et al 40 conducted a biomechanical study comparing unipedicular and bipedicular direct PMMA needle injection. Bipedicular injection of 10 mL (5 mL per side) improved vertebral strength significantly (P ≤ 0.05) more than a unipedicular injection of 6 mL; however, the unipedicular PMMA injection improved vertebral strength to a significantly (P ≤ 0.05) greater degree than that of the unfractured osteoporotic vertebrae. Both injection techniques r estored the stiffness of fractur ed vertebrae to prefracture values equally well. The change in temperature within and surrounding the vertebral body injected with PMMAhas been reported in cadaveric vertebrae maintained in a 37°C water bath. 41 Temperature recording was done at the anterior cortex, in the center of the vertebral body, and in the spinal canal. The two cements tested were Simplex P, a PMMA cement, and Orthocomp (Orthovita, Malvern, PA), an experimental non-PMMA cement. Both cements rose to a 41°C peak temperature in the spinal canal; within the vertebral body, Simplex P injection resulted in both a significantly (P ≤ 0.05) higher peak temperature and duration of heating >50°C compared with Orthocomp. The cooling effect of con- tinuous body fluid circulation, as would occur in vivo, has not yet been studied. Future Considerations Percutaneous injection of vertebral body pathology as a therapeutic technique is still in the investigation period, and appropriate indications for use need to be refined. Although multiple clinical series have documented the efficacy of both vertebroplasty and kyphoplasty in relieving pain from both vertebral metastases and osteoporotic fractures, no controlled clinical trials have documented their efficacy compared with other treatment modalities. 42 Current indications in patients with osteoporotic fractures include pr ogressive kyphosis, multiple fractures with collapse, incapacitating pain with either acute or subacute fractures requiring narcotic use, and chronic pain from vertebral fracture nonunion. The use of vertebral injection in unfractur ed vertebrae surrounding fractures remains controversial; a pr ospective random- ized series of patients with and without prophylactic adjacent injections is needed to determine the efficacy of this indication. Prophylactic use to prevent fracture in severely osteopen- ic patients currently has no clinical or scientific basis and thus is not indicated. In patients with metastatic disease with or without definite pathologic fracture, vertebral injection may provide pain relief and stability and may be an excellent adjuvant to radiation and chemotherapy. Prospective long- term series are needed to better un- derstand the biologic consequences of injection at the incident vertebral level and the effect on adjacent levels, which may become more prone to fracture. 43 Concurrent use with techniques such as radiofrequency abla- tion or stereotactic radiosurgery has yet to be investigated. The optimal technique for vertebral body augmentation is yet to be determined. Factors such as ease of use, efficacy, risks, complications, ability to correct deformity, and cost all must be considered when decid- ing whether vertebr oplasty or kyphoplasty is optimal in a particular clinical situation. Reported theoretical benefits of kyphoplasty, such as direct correction of collapse and deformity and fewer complications from the high-viscosity, low-pressur e injection, have yet to be demonstrated clinically in a direct comparison with postural correction obtained with vertebroplasty. The theoretical clinical benefits of restoration of vertebral height and kyphosis corr ection affect- ing future fractures, respiratory compromise, and pain have yet to be shown in long-term prospective clinical series. Cost factors include the vertebroplasty injection system or the inflatable bone tamp system, as well as costs associated with the use of in- terventional radiology or an operat- ing room and hospitalization. In patients with unremitting pain and minimal or no deformity, vertebroplasty may provide equally excellent clinical results with less cost than kyphoplasty and may obviate the need for general anesthesia. Kyphoplasty, however, may provide greater improvement of kyphotic collapse and Jeffrey M. Spivak, MD, and Michael G. Johnson, MD Vol 13, No 1, January/February 2005 15